You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@spamassassin.apache.org by "Daryl C. W. O'Shea" <sp...@dostech.ca> on 2006/09/30 22:22:08 UTC
Working on 3.1.6...
Hi folks,
I'd like to get a 3.1.6 tarball rolled by Tuesday so that we can hit yet
another release target. There are 16 bugs in the 3.1.6 queue. None of
the bugs needing review are complicated and most are quite trivial.
This is ready to go unless anyone has heard more from Joe:
- 5098 Add ecelerity received header format to the Rece...
These need 1 vote:
- 5111 FORGED_JUNO_RCVD FP
- 5105 M::SA::Client doesn't always catch failed connec...
- 5048 "spamassassin --lint" uses network rules, AWL
- 5040 'meta test has dependency with zero score' shoul...
- 5081 spamd_hup.t failures due to new kids being start...
- 5077 SPF plugin defaults to SOFTFAIL when DNS queries...
- 5044 sa-update can't update rules that depend on plug...
These need 2 votes:
- 5080 RCVD_ILLEGAL_IP doesn't hit as much as it should
- 4940 spamassassin fails to parse dates using obs-zone...
- 5089 Unable to add header with single digit zero
- 5094 Use of uninitialized value in pattern match (m//...
These will get kicked to 3.1.7 if nobody else comments on or takes care
of them soon:
- 4717 EnvelopeFrom option should be documented
- 4663 SpamAssassin can't find rules on Win32
- 4678 Two $'s in the MSG id causes 9.4 point in scoreset 2
- 4931 UnixNFSSafe.pm intermittently causes a race condition.
Please take some time to review at least the 11 needing votes if you
haven't already done so.
Thanks,
Daryl
Re: Working on 3.1.6...
Posted by "Daryl C. W. O'Shea" <sp...@dostech.ca>.
Daryl C. W. O'Shea wrote:
> Theo Van Dinter wrote:
>
>>> they're just config files, but the rules are a pretty big component
>>> of the product. In any case, there has only been 2 commits (one of
>>> which was the addition of a "1;") in the last 6+ months from all 5
>>> committers combined, so it's definitely not a pressing issue.
>>
>> Um. Where were you looking? I haven't been doing a ton of updates,
>> but it's
>> more than 2 in 6 months.
>
> I was referring to committers that aren't PMC members.
Bah, that's still not clear.
There have been 2 SVN commits total (not just to rule updates, but to
the entire codebase/sandboxes/etc) from the five people listed under
"Committers:" in the trunk CREDITS file in the last 6 months.
I realize you've done a number of rule updates in the same time.
I was just saying that it doesn't really matter (right now, anyway) if
committers (that aren't PMC members) have access to the zones box for
rule update purposes since none of them have been active (so I doubt
they'll be wanting to publish any rule updates).
Daryl
Re: Working on 3.1.6...
Posted by "Daryl C. W. O'Shea" <sp...@dostech.ca>.
Theo Van Dinter wrote:
>> they're just config files, but the rules are a pretty big component of
>> the product. In any case, there has only been 2 commits (one of which
>> was the addition of a "1;") in the last 6+ months from all 5 committers
>> combined, so it's definitely not a pressing issue.
>
> Um. Where were you looking? I haven't been doing a ton of updates, but it's
> more than 2 in 6 months.
I was referring to committers that aren't PMC members.
>> Anyway, the 3.1.6 queue is empty now, except for bug 5044. I just want
>> to make sure that you weren't reserving exit code 2 for anything before
>> I commit the sa-update lint changes.
>
> Geez, you guys cleared it out without me. <sniff>
>
> I think 2 is probably fine, though the description is incorrect (or misleading
> anyway):
>
> +An exit code of C<2> means that a lint check of the current site config failed.
>
> It means that linting of a channel failed. It could have been the site
> config, or it could have been the channel. There's no way to differentiate.
No, that's correct. It first lints just the local site config (and
exits with 2 if it fails). Then it does the channel updates, as before,
exiting with 4+ for whatever reason (lint failure, bad checksum, etc).
I've got it linting the site config by itself first so that we don't
waste mirror resources on installs that are never going to successfully
lint a channel update due to a broken local site config.
Daryl
Re: Working on 3.1.6...
Posted by Theo Van Dinter <fe...@apache.org>.
On Tue, Oct 03, 2006 at 08:24:10PM -0400, Daryl C. W. O'Shea wrote:
> As for who should have access to the zones box... (sort of related) I'm
> a little concerned about what kind of oversight we are required (by the
> ASF) to have over rule updates. If we need three PMC +1s for a release,
> what do we need for rule updates? I suppose you could argue that
As far as I know, the ASF has no opinion on the rule updates and so it's up to
the PMC. At last check, we decided that updates were CTR, and actually doing
an update doesn't require a vote -- especially since updates for trunk are
automatic.
> they're just config files, but the rules are a pretty big component of
> the product. In any case, there has only been 2 commits (one of which
> was the addition of a "1;") in the last 6+ months from all 5 committers
> combined, so it's definitely not a pressing issue.
Um. Where were you looking? I haven't been doing a ton of updates, but it's
more than 2 in 6 months.
> Anyway, the 3.1.6 queue is empty now, except for bug 5044. I just want
> to make sure that you weren't reserving exit code 2 for anything before
> I commit the sa-update lint changes.
Geez, you guys cleared it out without me. <sniff>
I think 2 is probably fine, though the description is incorrect (or misleading
anyway):
+An exit code of C<2> means that a lint check of the current site config failed.
It means that linting of a channel failed. It could have been the site
config, or it could have been the channel. There's no way to differentiate.
--
Randomly Selected Tagline:
You will be honored for contributing your time and skill to a worthy cause.
Re: Working on 3.1.6...
Posted by "Daryl C. W. O'Shea" <sp...@dostech.ca>.
Theo Van Dinter wrote:
> On Tue, Oct 03, 2006 at 07:09:48PM -0400, Daryl C. W. O'Shea wrote:
>> Cool, thanks. BTW, we need to put out a new 3.1 rule update before
>> 3.1.6 is released (due to eval test code being removed). I'm assuming
>> you need to do that (or are you automatically rsyncing to your box?).
>> I'm not quite sure how that's setup as I don't have an account on the
>> zone box.
>
> So right now what happens is that the updates get written on the zone box,
> and I have an rsync cronjob running every 15m to copy it to my machine
> which serves it out. I put up a wiki doc to explain how to do the manual
> (3.1) updates (http://wiki.apache.org/spamassassin/ManualRuleUpdates),
> but you would need an account on the zones machine. We ought to add that
> in as a step to be done -- I'm not sure if we want to restrict that to
> PMC or committer or what. :|
Yeah, I saw the wiki page (pretty similar to what I'm doing myself), I
just wasn't sure how updates were handled on your mirror since nobody
else has done an update yet.
As for who should have access to the zones box... (sort of related) I'm
a little concerned about what kind of oversight we are required (by the
ASF) to have over rule updates. If we need three PMC +1s for a release,
what do we need for rule updates? I suppose you could argue that
they're just config files, but the rules are a pretty big component of
the product. In any case, there has only been 2 commits (one of which
was the addition of a "1;") in the last 6+ months from all 5 committers
combined, so it's definitely not a pressing issue.
Anyway, the 3.1.6 queue is empty now, except for bug 5044. I just want
to make sure that you weren't reserving exit code 2 for anything before
I commit the sa-update lint changes.
Daryl
Re: Working on 3.1.6...
Posted by Theo Van Dinter <fe...@apache.org>.
On Tue, Oct 03, 2006 at 07:09:48PM -0400, Daryl C. W. O'Shea wrote:
> Cool, thanks. BTW, we need to put out a new 3.1 rule update before
> 3.1.6 is released (due to eval test code being removed). I'm assuming
> you need to do that (or are you automatically rsyncing to your box?).
> I'm not quite sure how that's setup as I don't have an account on the
> zone box.
So right now what happens is that the updates get written on the zone box,
and I have an rsync cronjob running every 15m to copy it to my machine
which serves it out. I put up a wiki doc to explain how to do the manual
(3.1) updates (http://wiki.apache.org/spamassassin/ManualRuleUpdates),
but you would need an account on the zones machine. We ought to add that
in as a step to be done -- I'm not sure if we want to restrict that to
PMC or committer or what. :|
--
Randomly Selected Tagline:
"But you have to allow a little for the desire to evangelize when you
think you have good news." - Larry Wall
Re: Working on 3.1.6...
Posted by "Daryl C. W. O'Shea" <sp...@dostech.ca>.
Theo Van Dinter wrote:
> On Tue, Oct 03, 2006 at 01:12:01PM -0400, Daryl C. W. O'Shea wrote:
>> Anyone? Still looking for 12 votes on 11 bugs...
>
> Thanks for taking point on this, btw. :)
No problem.
> And sorry I haven't poked at stuff yet. I've been trying to shift focus
> at work to actually devoting the 20% time I'm supposed to be doing,
> not the 40-50% I was doing for a few months.
Yeah, that's probably for the best. Too much for too long could
suddenly turn in to none at all. :(
> Anyway, assuming I'll prod around this afternoon/evening and see what I can
> do. :)
Cool, thanks. BTW, we need to put out a new 3.1 rule update before
3.1.6 is released (due to eval test code being removed). I'm assuming
you need to do that (or are you automatically rsyncing to your box?).
I'm not quite sure how that's setup as I don't have an account on the
zone box.
Daryl
Re: Working on 3.1.6...
Posted by Theo Van Dinter <fe...@apache.org>.
On Tue, Oct 03, 2006 at 01:12:01PM -0400, Daryl C. W. O'Shea wrote:
> Anyone? Still looking for 12 votes on 11 bugs...
Thanks for taking point on this, btw. :)
And sorry I haven't poked at stuff yet. I've been trying to shift focus
at work to actually devoting the 20% time I'm supposed to be doing,
not the 40-50% I was doing for a few months.
Anyway, assuming I'll prod around this afternoon/evening and see what I can
do. :)
--
Randomly Selected Tagline:
"I don't get the army: they kick you out for being gay, but their big plan to
improve moral is a make-over." - Bill Maher
Re: Working on 3.1.6...
Posted by "Daryl C. W. O'Shea" <sp...@dostech.ca>.
Anyone? Still looking for 12 votes on 11 bugs...
Daryl C. W. O'Shea wrote:
> Hi folks,
>
> I'd like to get a 3.1.6 tarball rolled by Tuesday so that we can hit yet
> another release target. There are 16 bugs in the 3.1.6 queue. None of
> the bugs needing review are complicated and most are quite trivial.
> These need 1 vote:
>
> - 5111 FORGED_JUNO_RCVD FP
> - 5105 M::SA::Client doesn't always catch failed connec...
> - 5048 "spamassassin --lint" uses network rules, AWL
> - 5040 'meta test has dependency with zero score' shoul...
> - 5081 spamd_hup.t failures due to new kids being start...
> - 5077 SPF plugin defaults to SOFTFAIL when DNS queries...
> - 5044 sa-update can't update rules that depend on plug...
>
>
> These need 2 votes:
>
> - 5080 RCVD_ILLEGAL_IP doesn't hit as much as it should
> - 4940 spamassassin fails to parse dates using obs-zone...
> - 5089 Unable to add header with single digit zero
> - 5094 Use of uninitialized value in pattern match (m//...
> Please take some time to review at least the 11 needing votes if you
> haven't already done so.
>
>
> Thanks,
>
> Daryl
>